What BLS’s 2025 AI Case Study Says About Budget Analysts

What is the short answer to the BLS case study?

BLS’s 2025 case study says AI could speed budget review and help create presentation visualizations. It expects conversations about budget nuances and alternatives to remain important because those discussions involve people. The study pairs this task-level judgment with projected U.S. budget-analyst employment growth of 3.9 percent from 2023 to 2033. These are different claims: one concerns possible changes to work; the other is an occupation-level outlook. Neither establishes how widely employers adopted AI, measures an employment effect caused by AI, or estimates an individual analyst’s chance of job loss. “Incorporating AI impacts in BLS employment projections: occupational case studies” describes research based on information available to BLS staff in June 2024 as they prepared that projection cycle. Its conclusion is a bounded historical assessment, not a current account of workplace practice. For an analyst deciding what to do now, inspect recurring tasks in the actual role: repeatable review, checking figures and assumptions, and explaining tradeoffs. Gather local evidence before considering a larger move. If workplace rules permit, test one narrow workflow and verify its results. The projection percentage alone does not warrant panic, reassurance, or retraining.

Sources: Incorporating AI impacts in BLS employment projections: occupational case studies

What period and evidence did the 2025 case study actually cover?

The BLS article “Incorporating AI impacts in BLS employment projections: occupational case studies” was published on February 10, 2025. Its budget-analyst discussion, however, describes a much earlier information boundary: the Employment Projections program’s research reflected staff interpretation of information available in June 2024, as they finalized projections for 2023–33. Those dates answer different questions. February 2025 is when readers could see the account; June 2024 marks the evidence window behind the case; 2023–33 names the projection period being constructed. Treating these as interchangeable can make a historical judgment sound like a current survey of workplace practice.

The scope is also specific. BLS presents several occupational case studies to explain how its Employment Projections staff consider possible labor-market effects of developing technology. For budget analysts, the article connects a view about possible task changes with a national occupation-level employment projection. It is not a report tracking a sample of budget offices, employees, or AI deployments. It does not measure what share of analysts used a tool, how frequently they used one, or whether a particular employer changed staffing. The case is evidence about BLS’s reasoning process and conclusion for that projection cycle, rather than a census of current adoption.

BLS explicitly cautions that other interpretations and conclusions could be drawn from the information then available. It also describes AI as a new, dynamic technology, with new information about development and potential uses continuing to emerge. The article says the Employment Projections program would keep assessing possible employment impacts and make appropriate updates in future projection sets. This qualification matters: the case study is not a timeless verdict about what budget analysts can or cannot do. It records a conditional judgment under a stated evidence cutoff. Newer tools, rules, procurement decisions, or workplace practices may change the present-day task picture without changing what BLS said it knew in June 2024.

A careful reading therefore keeps the publication date, evidence cutoff, base year, and horizon in view before interpreting the headline result. The 2023–33 figure belongs to a national projection model, whose purpose is to estimate occupational employment over that period; it cannot tell an individual whether their own role is secure or exposed. Nor does the date of publication establish that employers had adopted the tools discussed. For a present decision, the article can provide a useful historical baseline: BLS reasoned about potential productivity in some tasks while considering continued demand for other work. Current labor context should come from newer evidence, and a worker’s next step should depend on the actual duties, policies, and constraints in their workplace. That approach preserves the historical conclusion while avoiding the mistake of turning it into a report it was not.

Sources: Incorporating AI impacts in BLS employment projections: occupational case studies

What did BLS say AI might change in budget analysts' work?

BLS's February 2025 article, “Incorporating AI impacts in BLS employment projections: occupational case studies,” describes a conditional expectation, not a report that budget offices had already adopted AI. For budget analysts, the authors expected software advances to speed parts of budget review and potentially provide data-visualization tools for presentations. They also expected conversations about the nuances and alternative paths in proposed budgets to remain difficult to replace. This is an expectation of uneven task change, not a claim that an analyst's whole role is automated or protected. The mechanism in BLS's reasoning begins with the occupation's existing reliance on computer software. If new software helps analysts process or present information, the same staffing may be able to complete some review work more productively. BLS then identifies communication and customer-service duties as a source of continuing demand: analysts discuss what a proposal means, including its tradeoffs and alternatives. Productivity in one part of the job does not mechanically determine employment: demand for the underlying work and how organizations use saved time also matter. This is projection logic, not a measured estimate of time saved or staffing needs. The article uses cautious language for a reason. “Will likely speed” and “will likely continue” describe BLS staff's judgment when preparing the 2023–33 projections. They do not report a controlled trial of a budget-analysis tool, an adoption survey of budget offices, or observed changes in analyst staffing caused by AI. BLS says the case studies reflect its interpretation of information available in June 2024, when the projections were being finalized, and acknowledges that other interpretations were possible as the technology and its uses changed. Read the passage as a dated forecast about possible task effects, not a live census of workplace practice. The example is also narrower than the occupation. It names budgeting review, visualizations, and stakeholder discussion; it does not say that AI can independently decide whether a proposal is lawful, affordable, or aligned with a program's goals. Nor does it document a particular system's accuracy, access to records, or ability to handle exceptions. Likewise, BLS's expectation that discussion remains human does not make every conversation immune to change: preparation, summaries, or visual materials around a meeting could change even if people still explain and debate the choices. BLS paired this reasoning with projected budget-analyst employment growth of 3.9 percent from 2023 to 2033, about as fast as the average for all occupations. That figure reflects the projection for the occupation after the agency considered its assumptions about productivity and demand. It does not show that AI caused growth, that adoption was widespread, or that every budget office would retain the same roles. Take the case seriously without treating the number as a personal forecast: identify which recurring review or presentation tasks might be assisted, then ask what verification and stakeholder work remains in the actual job. The case supports that task-level question; it cannot answer it for a specific employer.

Sources: Incorporating AI impacts in BLS employment projections: occupational case studies

Why is budget review only one part of the analyst's work?

BLS's occupational profile describes budget analysis as a connected set of duties, not a single act of checking a spreadsheet. Analysts work with program and project managers to develop budgets, review requests for completeness, accuracy, and compliance, and combine unit budgets into an organization-wide plan. They then monitor spending, estimate future needs, analyze program costs and benefits, recommend funding levels, and explain findings. The Occupational Outlook Handbook also says officials or executives usually make the final budget decision, relying on analysts to prepare information for it. A review is therefore one point in a longer chain. A request must be complete, accurate, and compliant; combined requests must fit a consolidated budget. After approval, spending is monitored against the plan. BLS describes these responsibilities but does not measure their time share or variation by employer. The chain also extends beyond arithmetic. BLS says analysts may use cost-benefit analysis to assess tradeoffs and alternative funding methods, examine past budgets and economic or financial developments, and recommend cutting or redistributing funds. They require attention to the program's purpose and the decision at hand; a total alone cannot settle comparability, permission, or which tradeoff to accept. The occupational profile establishes that analysts do this evaluative work; the distinctions here are an interpretation of how the listed duties connect, not a claim that every analyst makes every policy judgment. Communication is part of the work, too. The Handbook describes explaining funding requests within an organization, to legislators, and to the public. It notes that analysts may write reports on differences between approved budgets and actual spending, explain discrepancies, and recommend changes. In government roles, they may attend hearings; some may assess policy or draft budget-related legislation. A chart or concise summary can make a pattern easier to see, but the recommendation still has to be intelligible to the people deciding or scrutinizing it. O*NET's Budget Analysts record supplies a structured occupational task and activity inventory that can help readers check whether the general profile resembles their own work. Like the Handbook, it describes an occupation in aggregate. Neither source tells an individual how their duties are divided, what software their office permits, which records are restricted, or who signs off on a recommendation. The sources are useful for naming the components of the job, not for assigning a universal share of work to each component or a single exposure label to everyone with the title. For interpreting the BLS AI case study, this wider bundle changes the question. Instead of asking whether budget analysts are exposed as one block, separate repeatable record handling and review from data lineage, rule interpretation, analysis of alternatives, recommendation authority, and explanation. Those steps can depend on one another: a polished output is useful only if its inputs and categories make sense in context and a responsible person can explain what follows from it. That is why the case study's focus on faster review does not describe the whole occupation. The proper unit for a reader's assessment is the actual task and its dependencies, checked against local rules and workflow, rather than the job title alone.

Sources: Budget Analysts; Budget Analysts

How can a faster review create new checking work?

A faster first pass is useful only when the analyst can establish that figures, categories, periods, and assumptions match source records and budget rules. A tool may shorten the time needed to assemble a summary or visualization; that does not establish that the analysis is accurate or fit to guide a funding decision. Review may reduce effort, shift effort toward checking, or yield little net saving if corrections are extensive. Measure the whole cycle, not just the first artifact. Imagine an analyst preparing a variance summary for a program manager. A permitted tool organizes expenditure lines and drafts a chart comparing planned and actual spending. Before using it, the analyst traces totals to source records, confirms categories mean the same thing in each period, and checks that dates and accounting cutoffs align. An apparent overrun may reflect timing, changed classification, or a substantive increase; each calls for a different explanation. This is illustrative, not a BLS-reported workflow or evidence that a tool performs these checks reliably. The checks matter because the budget analyst's task bundle includes more than assembling numbers. The Bureau of Labor Statistics’ Occupational Outlook Handbook says analysts review proposals for accuracy and compliance, monitor spending, estimate needs, analyze costs and benefits, and explain discrepancies while recommending changes. This occupation-level description does not measure time spent on each duty, but shows why a summary can sit within a chain of interpretation and communication. A chart may still compare unlike periods, omit a program change, or miss a rule limiting fund use. Suppose a draft suggests shifting funds between lines. The analyst verifies that records support the amount, checks whether the shift is allowed, and assesses whether it makes sense for program needs. Errors or unsupported assumptions add correction time. A manager may also need a defensible explanation. This does not prove that systems cannot assist with classification or rule lookup; a trial should specify tool access, checks performed, and who validates the output. The 2025 BLS Monthly Labor Review case study anticipated that AI might speed budget review and support data visualization, while describing nuanced discussion of alternatives as likely to continue through human conversations. This is a projection-era rationale, not a controlled test or measured finding about current review time. Plausible capability is not adoption, and a faster draft is not proof of higher productivity or fewer analyst hours. For an approved trial, compare assisted and ordinary processes on a similar task, recording corrections. Track total review time, changes required, traceability of figures and assumptions, and clarity for the reader. Count checking alongside drafting. If a quicker first pass creates a longer or less reliable final review, it has not shown a net gain. If it reduces routine assembly while preserving validation, it may help. Either result describes that workflow, not the occupation or future staffing.

Sources: Budget Analysts; Incorporating AI impacts in BLS employment projections: occupational case studies

Where does accountability remain when analysis is assisted?

The BLS case study does not say that an AI system takes responsibility for a budget recommendation. It describes review and possible presentation support alongside communication about budget nuances and alternatives, which BLS expected to continue through conversations between people. BLS’s Occupational Outlook Handbook says analysts prepare information for budget decisions, while government officials or private-company executives usually make those decisions. An assisted draft does not establish who checked it, who can approve a change, or who must explain it. That distinction is an interpretation of BLS’s described duties, not a universal legal rule. A budget process can involve assembling figures, checking a request for accuracy and compliance, comparing costs and benefits, recommending a funding level, and explaining the recommendation to managers, legislators, or the public. The Handbook lists these as parts of budget analysts’ work. They should not be collapsed into whether a tool can produce a summary. A system might assist with an information product while the organization still needs a process for validating inputs, resolving exceptions, and checking the proposed use against applicable rules. The case study’s emphasis on discussion matters because a recommendation can require people to examine what a budget change means for a program and what alternatives are available. BLS treats those conversations as difficult to replace; it does not establish that every exchange is beyond technical assistance or that every analyst has the same authority. Producing analysis, authorizing a decision, and answering for a recommendation are related but distinct responsibilities. CBO’s report, “How CBO Uses Artificial Intelligence Tools to Increase Efficiency,” offers a federal example. CBO says its staff investigated tools while recognizing limits and risks, and that tools can assist certain tasks without replacing analysts’ professional expertise and judgment. This is CBO’s account of its own work, not proof of common practice across budget offices. For a workflow map, ask: Who confirms the source data and period? Who checks classifications, assumptions, and compliance? Who may recommend or approve a funding change? Who answers stakeholders when numbers and program context do not align neatly? Recording these handoffs can show where assistance might fit and where review or communication remains necessary. It does not mean tasks involving people are immune to change; an organization could redesign who performs parts of the process. It means a faster draft alone cannot tell you where decision authority sits or how responsibility is assigned.

Sources: Budget Analysts; Incorporating AI impacts in BLS employment projections: occupational case studies; How CBO Uses Artificial Intelligence Tools to Increase Efficiency

How should the 3.9% figure be read beside BLS's newer outlook?

The 3.9 percent figure and the newer 2 percent figure answer related but different questions. In “Incorporating AI impacts in BLS employment projections: occupational case studies,” the Bureau of Labor Statistics (BLS) projected that U.S. budget-analyst employment would rise 3.9 percent between 2023 and 2033. That was part of the 2023–33 projection cycle, informed by Employment Projections staff’s interpretation of information available in June 2024. The current “Budget Analysts” profile projects 2 percent growth from 2025 to 2035. It estimates about 2,900 annual openings, most from workers transferring or leaving the labor force. These national projections cover different base years and horizons. It is tempting to subtract the newer rate from the older one and call the difference an AI effect. That conclusion would go beyond what BLS reports. Different assumptions, data, economic contexts, and horizons inform each cycle. The change does not isolate AI’s contribution, establish faster adoption, or show that an office plans to reduce staff. The 3.9 percent figure is not a measured result of AI use; it describes projected net employment change across the occupation. Openings also need a careful reading. The 2025–35 profile’s estimate of roughly 2,900 annual openings is not a claim that employment will expand by that many positions each year. BLS says most openings are expected from replacement needs, while the occupation’s total employment is projected to grow modestly. Openings are projected opportunities across the national occupation, not a guarantee of a vacancy in a particular region, agency, company, or specialty. BLS’s own methodological cautions help explain why task change and employment change need not move in lockstep. The 2025–35 “Employment Projections and Occupational Outlook Handbook News Release” says that projections involve inherent uncertainty and advises attention to the direction and relative size of changes rather than treating precise values as certain. It also says adjustments for emerging technologies are generally applied conservatively, such as when convincing evidence indicates a long-term structural change. In the 2025 case-study article, BLS explains that new technology can alter the composition or weighting of an occupation’s tasks without necessarily changing total employment. The case study is a bounded judgment about potential employment impacts, not a census of current tool use or time saved. This is useful counterevidence to two opposite shortcuts. Projected growth does not show that AI leaves the work untouched: tools may change review, preparation, or the mix of tasks even while overall employment rises. But possible productivity gains do not, by themselves, establish that AI will eliminate budget-analyst demand. An occupation-level projection cannot resolve one office’s staffing choice. BLS’s 2025–35 occupation profile is the better source here for current national labor context; the older case study remains useful for understanding how BLS reasoned about AI during the earlier cycle. For a practical decision, keep these sources in their lanes. Use the 3.9 percent result to understand the historical projection and its assumptions. Use the 2 percent outlook and openings estimate as a current national reference, with the uncertainty BLS states. Then look for local evidence that neither figure supplies: whether your employer has approved tools, which duties are changing, what review standards apply, whether staffing or reporting responsibilities have shifted, and what skills the role now requires. Those facts can justify a small workflow experiment or a focused skill upgrade. The national percentages alone cannot tell you whether to leave budgeting, retrain, or expect a particular outcome at work.

Sources: Incorporating AI impacts in BLS employment projections: occupational case studies; Budget Analysts; Employment Projections and Occupational Outlook Handbook News Release

What does a real federal budget-office example add—and not add?

CBO's 2025 disclosure adds what the BLS case study cannot: an account of one federal budget institution's own inquiry into AI tools. How CBO Uses Artificial Intelligence Tools to Increase Efficiency, published May 12, 2025, says staff were investigating whether tools could improve efficiency and effectiveness while recognizing current limits and risks. CBO describes its work as gathering information, analyzing it, and clearly communicating results for Congress. It says AI can assist certain tasks but cannot replace analysts or their professional expertise and judgment. This establishes CBO's stated activity and position; it does not show that every budget analyst uses AI or that a measured productivity gain occurred.

The account complements but differs from Incorporating AI Impacts in BLS Employment Projections: Occupational Case Studies. BLS explains how Employment Projections staff interpreted information available in June 2024 while preparing national projections for 2023–33. Its budget-analyst discussion is a forward-looking judgment about possible task change and employment demand. CBO, by contrast, describes one agency's inquiry into tools for its work. One source concerns a projection method and occupational outlook; the other is an institutional snapshot. Neither confirms the other: CBO's activity does not validate the BLS forecast, and BLS's projection does not tell us what tools CBO or another office has approved or adopted.

CBO's account has limits. Its mission, data, procurement rules, and review processes may differ from those of state and local offices, nonprofits, or private employers. The letter does not quantify time saved, compare output quality, or report staffing effects. Its statement that tools can assist certain tasks does not show that all tasks will change; its emphasis on expertise does not show that every task will remain unchanged. It supports a narrower conclusion: CBO saw potential for assistance, acknowledged risks, and considered professional judgment important to its analysts' work.

For an analyst deciding what to do, treat this as a prompt for local questions, not a market-wide adoption signal. Does your organization permit a particular tool with the data involved? Which uses are approved, and what must be checked before work reaches a decision-maker? Who verifies figures and assumptions, and who explains the recommendation? Clear rules may support a small, supervised trial on a suitable recurring task; restrictions may make policy knowledge or a better non-AI workflow the immediate priority. These are practical implications of comparing the sources, not recommendations issued by BLS or CBO. The evidence does not support a personal job-loss estimate. It supports checking workplace rules and workflows before treating one federal example as reason to retrain, change roles, or assume adoption is widespread.

Sources: How CBO Uses Artificial Intelligence Tools to Increase Efficiency; Incorporating AI impacts in BLS employment projections: occupational case studies

Which parts of your own work should you map first?

Start with the recurring work, not the job title. O*NET OnLine’s Budget Analysts profile is a useful prompt: it lists examining estimates for accuracy and compliance, analyzing reports, monitoring spending, preparing recommendations, interpreting directives, and communicating budgets to stakeholders. The Bureau of Labor Statistics’ Occupational Outlook Handbook likewise describes combining budgets, tracking spending, estimating needs, comparing costs and benefits, and explaining recommendations. These are occupation-level descriptions. They can help you remember duties, but they do not tell you how your office divides the work, which systems you use, or which decisions you personally own. Replace the generic list with a recent work cycle from your own desk. For each recurring task, write down its input, expected output, frequency, and approximate time. Then add the cost of an error, the checks required before anyone relies on the result, who has authority to approve or act on it, who needs an explanation, and whether workplace policy permits a tool to handle the data. This is a practical mapping framework, not a validated score. Its purpose is to distinguish tasks that sound similar but carry different responsibilities. Drafting a summary from an approved, clean table is not the same task as deciding whether a variance is permissible or recommending a funding change. Sort the entries into three working groups. First, identify repeatable information handling: compiling records, arranging figures, preparing a first draft, or turning approved data into a visualization. Second, mark validation and interpretation: tracing figures to source records, checking periods and categories, applying directives, assessing assumptions, and deciding what a difference means. Third, note explanation and negotiation: clarifying a recommendation with program managers, responding to questions, or discussing alternatives with people responsible for budgets. The groups are a way to organize your own observations, not a claim that every office follows one workflow. A single assignment may move through all three. Finally, mark constraints that can outweigh apparent technical fit: restricted or sensitive data, lack of approved system access, rules about where records may be processed, high correction costs, or a reviewer who must sign off. The Congressional Budget Office’s account of its own AI use is a reminder to consider institutional limits and review alongside possible assistance; it is not evidence that your employer has adopted similar tools. A useful map might therefore point to no experiment yet, or to one low-stakes task that policy allows you to test under supervision. Either result is more actionable than treating BLS’s national case study as a personal risk measure.

Sources: Budget Analysts; Budget Analysts; How CBO Uses Artificial Intelligence Tools to Increase Efficiency

What is a proportionate next move for a budget analyst?

Begin with one recurring task . The BLS case study is a reason to examine review and visualization work, not to buy a credential or leave budgeting because of a projection. If permitted, test with approved data and a reviewer. Compare ordinary and assisted workflows: Were figures and categories correct? How many corrections were needed? Did the full cycle take less time, including checks? Was it clear, and who was responsible? A quicker draft does not prove a better workflow if review absorbs the time saved.

Let the comparison set your next step. For repeated data handling, learn the approved tool. If errors stem from mismatched periods, weak source trails, or unclear assumptions, strengthen data interpretation and checking. If explaining tradeoffs is difficult, practice concise recommendations grounded in the figures. These skills travel across tools; interfaces change.

An upgrade in budgeting may fit better than an adjacent move into finance or program analysis. Consider a larger transition only if your duties or goals support it. Weigh options against salary needs, location, credentials, study time, health, and family responsibilities. The BLS Occupational Outlook Handbook describes duties and national projections, but cannot settle local questions. The older 3.9% figure cannot answer them either. The free task checker organizes task-level change-pressure signals; it is not a validated probability of displacement. Observe one compliant workflow, then decide what evidence warrants a wider move.

Sources: Budget Analysts; Incorporating AI impacts in BLS employment projections: occupational case studies

Questions readers ask